Upload README.md with huggingface_hub

README.md

@@ -7,7 +7,7 @@ model_name: Gemma-4-26B-A4B-it-Cerebellum-v6-GGUF
 model_creator: google
 model_type: gemma4
 quantized_by: deucebucket
-pipeline_tag: text-generation
+pipeline_tag: image-text-to-text
 tags:
   - GGUF
   - gemma4
@@ -19,6 +19,8 @@ tags:
   - moe
   - 3-bit
   - conversational
+  - multimodal
+  - vision
 ---
 
 # Gemma 4 26B-A4B-it — Cerebellum v6 GGUF
@@ -29,12 +31,15 @@ Cerebellum v6 is an ablation-guided mixed-precision GGUF quantization of [google
 
 This is a 26B-parameter MoE model with 4B active parameters per token, 128 experts per layer, and 30 layers. This release uses tensor-level precision overrides selected from 140+ ablation experiments across six internal iterations, including per-layer MoE router surgery.
 
+**This model supports vision** when used with the included mmproj file. See [Vision Support](#vision-support) below.
+
 ## At a Glance
 
 | | |
 |---|---|
 | **File** | `gemma-4-26B-A4B-it-cerebellum-v6.gguf` |
-| **Size** | 11.7 GB |
+| **mmproj** | `mmproj-google_gemma-4-26B-A4B-it-f16.gguf` |
+| **Size** | 11.7 GB (backbone) + 1.14 GB (mmproj) |
 | **Base model** | `google/gemma-4-26B-A4B-it` |
 | **Base quant** | Q3_K_M with [bartowski's imatrix](https://huggingface.co/bartowski/google_gemma-4-26B-A4B-it-GGUF) |
 | **Format** | GGUF, mixed precision |
@@ -144,13 +149,90 @@ Q8_0 is the only precision that improves PPL. K-quant formats (Q6_K, Q2_K) use 2
 
 91 tensor-level overrides + 1 surgical router recast on top of Q3_K_M base.
 
+## Vision Support
+
+This model is multimodal — it can process images alongside text. Vision requires two files used together:
+
+- `gemma-4-26B-A4B-it-cerebellum-v6.gguf` — the text backbone (this file)
+- `mmproj-google_gemma-4-26B-A4B-it-f16.gguf` — the vision encoder + projector (1.14 GB)
+
+The image token `<|image|>` (token ID 258880) is already in the vocabulary. No metadata changes needed.
+
+### Usage with llama-server
+
+```bash
+llama-server \
+  -m gemma-4-26B-A4B-it-cerebellum-v6.gguf \
+  --mmproj mmproj-google_gemma-4-26B-A4B-it-f16.gguf \
+  --jinja \
+  --reasoning off \
+  --reasoning-budget 0 \
+  -ngl 99 \
+  -c 4096
+```
+
+**Required flags:**
+- `--mmproj` — loads the vision encoder. The mmproj filename starts with `mmproj-` so it also works with `--mmproj-auto` auto-download if placed in the same directory.
+- `--jinja` — enables the Gemma 4 chat template (embedded in the GGUF; required for correct formatting)
+- `--reasoning off --reasoning-budget 0` — disables thinking mode which can cause infinite loops without dedicated reasoning tokens
+
+### Usage with curl
+
+```bash
+curl http://localhost:8080/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "gemma4-cerebellum",
+    "messages": [
+      {
+        "role": "user",
+        "content": [
+          {"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}},
+          {"type": "text", "text": "What is shown in this image?"}
+        ]
+      }
+    ]
+  }'
+```
+
+### Usage with ollama
+
+Create a `Modelfile`:
+
+```dockerfile
+FROM ./gemma-4-26B-A4B-it-cerebellum-v6.gguf
+FROM ./mmproj-google_gemma-4-26B-A4B-it-f16.gguf
+TEMPLATE {{ .Prompt }}
+```
+
+Then:
+```bash
+ollama create gemma4-cerebellum -f Modelfile
+ollama run gemma4-cerebellum
+```
+
+### Technical notes
+
+- The vision encoder has 27 layers with 1152 hidden dimension, projecting to 2816 (matching the text model's embedding dimension).
+- Image resolution: native input size 224×224, patches of 16×16. Dynamic resolution supported.
+- The mmproj was converted using llama.cpp's `convert_hf_to_gguf.py` from the original Google model and is redistributed under the Apache 2.0 license. Conversion credit: [bartowski](https://huggingface.co/bartowski).
+- Vision works out of the box — no special tokens, metadata edits, or re-quantization needed.
+
 ## Usage
 
+### llama.cpp
+
 ```bash
-# llama.cpp
+# Text only
 ./llama-server -m gemma-4-26B-A4B-it-cerebellum-v6.gguf -ngl 99 -c 4096
 
-# ollama
+# With vision
+./llama-server -m gemma-4-26B-A4B-it-cerebellum-v6.gguf --mmproj mmproj-google_gemma-4-26B-A4B-it-f16.gguf --jinja --reasoning off --reasoning-budget 0 -ngl 99 -c 4096
+```
+
+### ollama
+
+```bash
 ollama create gemma4-cerebellum -f Modelfile
 ollama run gemma4-cerebellum
 ```
@@ -164,9 +246,11 @@ Fits in 24 GB VRAM at full GPU offload with room for 4K context.
 - **Ablation experiments**: 140+ across 6 iterations (including 30-layer router surgery)
 - **Quantizer**: llama.cpp `llama-quantize` with `--tensor-type-file` overrides + `gguf_tensor_surgery.py` for router recast
 - **Hardware**: RTX 3090 (24 GB VRAM)
+- **Vision**: See [Vision Support](#vision-support) for details
 
 ## Credits
 
 - **Base model**: [Google Gemma Team](https://huggingface.co/google/gemma-4-26B-A4B-it)
 - **Imatrix**: [bartowski](https://huggingface.co/bartowski/google_gemma-4-26B-A4B-it-GGUF)
+- **mmproj conversion**: [bartowski](https://huggingface.co/bartowski/google_gemma-4-26B-A4B-it-GGUF)
 - **Method & quantization**: [deucebucket/osmosis](https://github.com/deucebucket/osmosis) — Cerebellum pipeline